Category talk:Mandelbrot sets

Discussion moved here

There has been some discussion elsewhere related to the structure of subcategories of this category. Because my reply raises many different issues, I'm moving the discussion here.

First, some background info…

The current subcat structure here is:

Mandelbrot sets (detail)‎ (2 C, 6 F) Misiurewicz point‎ (9 F) Near-copies of the Mandelbrot set within itself‎ (30 F) Mandelbrot sets (whole)‎ (92 F) Mandelbrot set animations‎ (26 F) Mandelbrot set derivatives‎ (3 C, 70 F) Buddhabrots‎ (1 P, 43 F) Mandelbulbs‎ (1 P, 23 F) Multibrot sets‎ (34 F)

Mandelbrot sets in grayscale‎ (27 F) Mandelbrot set interior‎ (21 F) Mandelbrot set shapes in other fractals‎ (30 F) Mandelbrot sets (zoom)‎ (empty) Mandelbrot set properties‎ (43 F) Mandelbrot set non-zoom sequences‎ (1 C, 22 F) Complex color plots of Mandelbrot iterations‎ (21 F) Mandelbrot set zoom sequences (static)‎ (90 F) Mandelbrot set videos‎ (6 F)

Category:Mandelbrot sets (detail)‎ was created (by me) in order to depopulate Category:Mandelbrot sets (zoom)‎, which I thought had an ambiguous/misleading name. User:Adam majewski, who created that category, agreed and suggested:

"I have thought about creating subcategories with respect to :

• image type ( static image, series of images, animation, videos)
• set  : whole set / detail of the set ( = zoom)
• algorithm ( DEM, ETM, ....)
• color ( 24 / 8/ 2 bit)
• part of plane : exterior / boundary / interior
• map ( quadratic/higher power / ... )

So every image could be categorized using each above subcategory."

My response is…

Sounds logical. However, I would say that the first image type, "[stand-alone] static image", doesn't warrant a special category, since that type constitutes the vast majority of both what we have and what people are looking for (i.e., the existence of that subcat would not appreciably improve the ability of an end-user to actually find something they're looking for). Likewise with "24-bit color depth", "quadratic map", "boundary", and—now that I think about it—probably "detail of set", as well. Would the existence of such subcats actually help people find stuff more easily? Probably not. OTOH, the "opposite" of each of these (series of images / animations / videos; reduced color depth; not the classic "z²+c" map; exterior-only or interior-only [do we even have any of these?]; whole set visible) are likely to be useful as subcats, since they will provide more focused collections of files for users to look through.

On a more prosaic matter, perhaps "mosaic" might be added as another "image type" in the list above. - dcljr (talk) 03:49, 11 July 2014 (UTC)

General categorizing principles

The discussion in the previous section raises the fundamental question of which of two possible approaches we will take in this category (called the "main category" in what follows):

1. Assume a "default" type of file to appear in the main category and only subcategorize based on distinguishing (non-default) characteristics.
2. Create multiple "partitions" (mutually exclusive and collectively exhaustive sets of subcategories, as suggested by Adam's list above), keeping the main category relatively free of files.

I tend to lean towards Option 1, but in this case I'm not so sure. I can see pros and cons to each approach…

Considerations in favor of Option 1:

• surely the most common approach here at Commons (since categories tend to be created without a larger substructure in mind, and only later is it recognized that subcats are even necessary), and hence this matches what users somewhat familar with Commons may be expecting
• each file would belong in fewer subcats, perhaps making it easier to properly/fully categorize new files?
• fewer subcats may make it easier to categorize files using existing tools (e.g., HotCat), depending on how the subcats are named

Considerations against Option 1:

• systematically encourages images to collect in the main category (until appropriate new subcats are identified)
• some users might be confused by the "missing" subcats
• users might not be able to find files they're looking for because each file would belong in fewer places
• what is considered the "default" file type is debatable (my definition: a stand-alone, static image of part^* of the classic quadratic-map Mandelbrot set, with black or other dark monochrome interior and multicolored exterior, colored to a 24-bit depth)

* I don't deny that many users will come looking for the full set, but of the two "scopes" (whole vs. part), the vast majority of our images are of just part of the set.

Considerations in favor of Option 2:

• more "logical" and appeals to categorizers' tendency to categorize [g]
• more subcats makes for more homogenous groupings (thus easier for users who already know what they want)
• users can expand little "category trees" on the Category page itself, so things are not as easily "lost" (as they used to be) by moving them deeper into the category structure
• enables more fine-grained category-intersection searches (for those users who know about and use those tools—not a large percentage of end-users!)

Considerations against Option 2:

• some subcats would be nearly as large as the main category would be if the subcat didn't exist (because some characteristics are shared by almost all of the files)
• this type of "exhaustive categorizing" would work better if categories acted like "tags" (i.e., if it were encouraged to mark images with all applicable general characteristics without worrying about how large the resulting categories got)
• some end-users may be confused by the existence of so many overlapping subcats, especially if they're looking for the typical "default" image type
• more subcats means more ambiguous cases, resulting in even more subcats (e.g., "Mandelbrot sets using indeterminate coloring schemes"?)

What say you? - dcljr (talk) 03:49, 11 July 2014 (UTC)

option 2 gives explicit categorisation, option 1 maybe unclear to user. First option is in Category:Julia sets. It is divided to connected an not connected (I did not know what to do with category : Julia Mandelbrot relationship map‎). BTW I think also about creating category Cremer julia sets ( inside connected). It can be empty because there is no images of such Julia sets, but it can be also good info - someone should do it (:-) --Adam majewski (talk) 08:07, 13 July 2014 (UTC)

Specific changes to current subcategories

In addition to the large issues outlined above, I am rethinking my creation of Category:Mandelbrot sets (detail)‎, the removal of which would necessitate the promotion of Category:Misiurewicz point‎ and Category:Near-copies of the Mandelbrot set within itself‎ to the main category level. Also, I want to rename Category:Mandelbrot set derivatives‎ to avoid using the suggestive term "derivatives". I would call it Category:Mandelbrot set generalizations, but the Buddhabrot is not really a "generalization" of the set (depending on how you define the word) so much as just a different coloring of it. (So, I guess Category:Buddhabrots‎ also belongs in the main category, for now.)

Furthermore, I'm not sure what the existence of Category:Complex color plots of Mandelbrot iterations actually adds to the situation. Aren't all (colored) images of the Mandelbrot set "Complex color plots"? If so, then this subcat should perhaps be merged into the parent Category:Mandelbrot set non-zoom sequences‎. (Actually, it might be a good idea to just rename the latter subcat to include the idea of "iteration(s)", since all the current content is related to that, anyway.) - dcljr (talk) 03:49, 11 July 2014 (UTC)

Specific file categorization issues

And then (finally) there's the whole issue of certain images in the category that are reportedly generated using "Mandelbrot's method" (whatever that means, exactly) but otherwise do not appear to be images of "the Mandelbrot set" (usually this is because they use a different mapping). A while back, I moved many of these to Category:Exponential maps and/or Category:Mandelbrot set shapes in other fractals (depending on what they looked like), but this is really getting beyond the limits of my familiarity with the subject matter (since I've never formally studied the topic), so I am unsure how to proceed. Take, for example, these three files:

• File:MANDEL Z-2+C medium.jpg (still in this category)
• File:Mandel SIN(Z)+C medium.jpg (ditto)
• File:MANDEL EXP(Z C5) medium.jpg (moved to Category:Exponential maps)

Where would you put them? - dcljr (talk) 03:49, 11 July 2014 (UTC)